Identification of ferroptosis-related genes in type 2 diabetes mellitus based on machine learning

Sen Wang; Yongpan Lu; Tingting Chi; Yixin Zhang; Yuli Zhao; Huimin Guo; Li Feng

doi:10.1002/iid3.1036

Identification of ferroptosis-related genes in type 2 diabetes mellitus based on machine learning

Immun Inflamm Dis. 2023 Oct;11(10):e1036. doi: 10.1002/iid3.1036.

Authors

Sen Wang¹, Yongpan Lu², Tingting Chi³, Yixin Zhang¹, Yuli Zhao¹, Huimin Guo¹, Li Feng¹

Affiliations

¹ Department of Medical Ultrasound, Shandong Medicine and Health Key Laboratory of Abdominal Medical Imaging, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qian Foshan Hospital, Shandong First Medical University, Jinan, Shandong, China.
² Department of Plastic Surgery, The First Clinical Medical College, Shandong University of Traditional Chinese Medicine, The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qian Foshan Hospital, Jinan, Shandong, China.
³ Department of Acupuncture and Rehabilitation, The Affiliated Qingdao Hai Ci Hospital of Qingdao University (West Hospital Area), Qingdao, Shandong, China.

Abstract

Background: Type 2 diabetes mellitus (T2DM), which has a high incidence and several harmful consequences, poses a severe danger to human health. Research on the function of ferroptosis in T2DM is increasing. This study uses bioinformatics techniques identify new diagnostic T2DM biomarkers associated with ferroptosis.

Methods: To identify ferroptosis-related genes (FRGs) that are differentially expressed between T2DM patients and healthy individuals, we first obtained T2DM sequencing data and FRGs from the Gene Expression Omnibus (GEO) database and FerrDb database. Then, drug-gene interaction networks and competitive endogenous RNA (ceRNA) networks linked to the marker genes were built after marker genes were filtered by two machine learning algorithms (LASSO and SVM-RFE algorithms). Finally, to confirm the expression of marker genes, the GSE76895 dataset was utilized. The protein and RNA expression of some marker genes in T2DM and nondiabetic tissues was also examined by Western blotting, immunohistochemistry (IHC), immunofluorescence (IF) and quantitative real-time PCR (qRT-PCR).

Results: We obtained 58 differentially expressed genes (DEGs) associated with ferroptosis. GO and KEGG enrichment analyses showed that these DEGs were significantly enriched in hypoxia and ferroptosis. Subsequently, eight marker genes (SCD, CD44, HIF1A, BCAT2, MTF1, HILPDA, NR1D2, and MYCN) were screened by LASSO and SVM-RFE machine learning algorithms, and a model was constructed based on these eight genes. This model also has high diagnostic power. In addition, based on these eight genes, we obtained 48 drugs and constructed a complex ceRNA network map. Finally, Western blotting, IHC, IF, and qRT-PCR results of clinical samples further confirmed the results of public databases.

Conclusions: The diagnosis and aetiology of T2DM can be greatly aided by eight FRGs, providing novel therapeutic avenues.

Keywords: bioinformatics; diagnostic; ferroptosis; gene expression omnibus; machine learning; type 2 diabetes mellitus.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Algorithms
Diabetes Mellitus, Type 2* / diagnosis
Diabetes Mellitus, Type 2* / genetics
Ferroptosis* / genetics
Humans
Machine Learning
RNA

Substances

RNA